AnandTech Storage Bench - The Destroyer

Our AnandTech Storage Bench tests are traces (recordings) of real-world IO patterns that are replayed onto the drives under test. The Destroyer is the longest and most difficult phase of our consumer SSD test suite. For more details, please see the overview of our 2021 Consumer SSD Benchmark Suite.

ATSB The Destroyer
Average Data Rate
Average Latency Average Read Latency Average Write Latency
99th Percentile Latency 99th Percentile Read Latency 99th Percentile Write Latency
Energy Usage

For SATA drives, the Samsung 870 EVOs turn in class-leading scores on almost all of the performance metrics. But these improvements are all marginal at best; the SATA interface bottleneck almost completely levels the playing field. The small improvements to read latency brought by the 870 EVO pale in comparison to what is achieved by even entry-level NVMe SSDs.

In stark contrast to the performance numbers, the 870 EVOs turn out to be the most power-hungry TLC drives in this bunch: they sacrifice some of the efficiency improvements the 860 EVO provided, even though drives like the SK hynix Gold S31 have been able to deliver significant improvement on this.

AnandTech Storage Bench - Heavy

The ATSB Heavy test is much shorter overall than The Destroyer, but is still fairly write-intensive. We run this test twice: first on a mostly-empty drive, and again on a completely full drive to show the worst-case performance.

ATSB Heavy
Average Data Rate
Average Latency Average Read Latency Average Write Latency
99th Percentile Latency 99th Percentile Read Latency 99th Percentile Write Latency
Energy Usage

The scores for the Heavy test paint much the same picture as for The Destroyer. The full-drive test runs additionally show that the worst-case performance of the mainstream SATA SSDs is still superior to many entry-level NVMe SSDs, even though the NVMe SSDs significantly outperform SATA for any more normal workload.

AnandTech Storage Bench - Light

The ATSB Light test represents ordinary everyday usage that doesn't put much strain on a SSD. Low queue depths, short bursts of IO and a short overall test duration mean this should be easy for any SSD. But running it a second time on a full drive shows how even storage-light workloads can be affected by SSD performance degradation.

ATSB Light
Average Data Rate
Average Latency Average Read Latency Average Write Latency
99th Percentile Latency 99th Percentile Read Latency 99th Percentile Write Latency
Energy Usage

On the Light test, the measurable but imperceptible performance advantages of the 870 EVOs over other SATA drives have basically disappeared. The read latency scores on the full-drive test runs may be a tiny bit better than the 860 EVO, but the only scores that have clearly shifted with this new generation are the energy consumption figures that have creeped up.

PCMark 10 Storage Benchmarks

The PCMark 10 Storage benchmarks are IO trace based tests similar to our own ATSB tests. For more details, please see the overview of our 2021 Consumer SSD Benchmark Suite.

PCMark 10 Storage Traces
Full System Drive Overall Score Average Bandwidth Average Latency
Quick System Drive Overall Score Average Bandwidth Average Latency
Data Drive Overall Score Average Bandwidth Average Latency

The Full System Drive test from the PCMark 10 Storage suite shows a much wider spread of performance scores among SATA drives than our ATSB traces, but also a much smaller advantage for the NVMe drives. Judging by this test, the 870 EVO offers a small but real improvement to performance compared to earlier SATA drives. The 4TB 870 QVO also scores quite well since it benefits from the same controller and has enough SLC cache to almost match the performance of the 4TB 870 EVO.

The subset of tests included in the Quick System Drive and Data Drive benchmarks show a more level playing field among SATA SSDs, and a greater advantage for NVMe drives. Since we run these tests before the Full System Drive test, each drive is closer to its fresh out-of-the-box state, which helps these tests get closer to showing the theoretical peak performance of a drive.

Introduction Synthetic Tests: Basic IO Patterns
POST A COMMENT

139 Comments

View All Comments

  • Jorgp2 - Wednesday, February 17, 2021 - link

    Nah, it all comes down to EFI vs MBR.

    MBR systems just look for a boot partition and boot it, EFI actually stores the location on the motherboard flash.
    Reply
  • WaltC - Wednesday, February 17, 2021 - link

    Agreed...;) I might say "UEFI vs. Legacy," though. And much of it has to do with the knowledge, experience, and skill of the computer operator...! No question about that. (It's actually MBR vs. GPT--GPT is better. All of my SATA HDDs are formatted GPT, etc. I have no MBR-formatted drives.)

    The list of NVMe drives here is curious. Where is the 980 Pro from Samsung?--been selling for a while now--one of several PCIe4 NVMe drives available. Makes me question when this article was actually written....? It seems incomplete or out of date. Samsung is a chip company no longer in the business of making platter drives (last I looked, anyway-do they still sell the Spinpoints?), so it's natural for them to sell SSDs of varying types, sizes, and prices, imo.
    Reply
  • Billy Tallis - Wednesday, February 17, 2021 - link

    I deliberately chose not to include 980 PRO results in the graphs for this review, because that's a silly comparison to make against a SATA drive. But if you really care, you can use Bench: https://www.anandtech.com/bench/product/2724?vs=27... Reply
  • Gigaplex - Wednesday, February 17, 2021 - link

    A half-decent UEFI implementation will enumerate the bootloaders (or fall back to the default /EFI/Boot/bootx64.efi path) if the drive isn't configured in the motherboard. Reply
  • Billy Tallis - Thursday, February 18, 2021 - link

    Unfortunately, "half-decent" can't be taken for granted. The ASRock motherboard in the new SSD testbed won't look in the standard path for a bootloader, but it will happily boot any Windows bootloader it finds. Reply
  • Duncan Macdonald - Wednesday, February 17, 2021 - link

    SATA may be declining but is far from dead - many motherboards (and laptops) only have one NVMe slot - adding a SATA drive is far easier than replacing the NVMe drive when it is the system drive. Also NVMe drives bigger than 4TB are rare and expensive so anyone needing large storage capacity (over 4TB) has a choice between SATA or expensive NVMe. Reply
  • npz - Wednesday, February 17, 2021 - link

    My thoughts too. Until I can get a 4tb or more nvne m.2 drive priced the same and also not have it throttle under sustained i/o I'll be sticking with mix of mechanical and ssd sata drives and u..2 drines and m.2 placed into u.2 adapters for 2.5 bay cooling for those few hot nvme drives I can't place a tall heatsink over Reply
  • Gigaplex - Wednesday, February 17, 2021 - link

    If you're prepared to suffer from mechanical performance limitations, then you shouldn't be worried about sustained IO throttling. Reply
  • npz - Thursday, February 18, 2021 - link

    As I said, I have a *mix* of drives for different workloads. If I'm doing concurrent read and write from a source drive to a destination drive for large video files I'm not doing that on the mechanical drives and doing that on m.2 drives without proper cooling will throttle. Reply
  • Oxford Guy - Friday, February 19, 2021 - link

    Some motherboard brands also don't know how to design a motherboard properly, like Gigabyte.

    Drives like the Inland Performance Plus (Phison) don't fit in boards like the Z390 UD.
    Reply

Log in

Don't have an account? Sign up now